Investigation of the operating point influence on the low-engine-order excitation in vaneless radial turbines

Abstract

Aerodynamic excitation in turbomachines can induce synchronous forced response vibrations, which may pose a high risk due to High-Cycle-Fatigue (HCF) failure of the rotor when excited at resonance. In the case of vaneless radial turbines, which are often used in turbochargers, the excitation mainly occurs at the so-called Low Engine Orders (LEO) stemming from asymmetrical geometry of the volute. These machines are operated dynamically over a wide range of operating conditions, constantly altering the harmonic blade forces and, hence, the aerodynamic excitation. The current work aims at providing a thorough understanding of the mechanisms for the variation of the LEO excitation as a function of operating point. Therefore, numerical and experimental investigations are performed on three vaneless radial turbines of similar size. The numerical investigations are performed at numerous operating points, employing transient Computational Fluid Dynamics (CFD) simulations. An analysis of local excitation variations on the blade surface at changed operating conditions has shown that these are governed by the wheel’s inflow velocity triangle and the dynamic pressure. Incidence variations, in particular, have a crucial impact on the harmonic pressure distribution and may cause a non-linear forcing behavior between operating points. Besides, an increase of the dynamic pressure ideally corresponds to a proportional rise in the excitation. A deviation to this proportionality occurs due to flow phenomena, such as the tip clearance flow, that modulate the forcing field originating from the volute. These observations are found to be valid but with a distinctive effect at the different LEOs and test objects. Consequently, the shape of an excitation map, describing the forcing under different operating conditions, is unique for each resonance crossing. However, it can be related to the excitation variation mechanisms of the local forcing. The numerical results are largely supported by the experimental vibration data, measured by means of blade tip timing. This confirms that significant differences in the aerodynamic excitation, sometimes at a factor of more than two, are present due to a change of operating conditions as well as between the mistuned response of the blades at a single resonance crossing.